-
Associations between riffle development and aquatic
biotafollowing lowhead dam removal
Danielle R. Cook & S. Mažeika P. Sullivan
Received: 20 January 2018 /Accepted: 2 May 2018 /Published
online: 10 May 2018# The Author(s) 2018
Abstract Dam removal is an increasingly commonriver restoration
option, yet some of the mechanismsleading to ecological changes
remain unquantified. Weassessed relationships between riffle
structure and ben-thic macroinvertebrate and fish assemblages 2
yearsafter a lowhead dam removal in Ohio, USA. Hydrogeo-morphic,
water-chemistry, and biotic surveys were con-ducted at seven study
riffles at six time intervals fromspring 2014 through summer 2015.
The density anddiversity of macroinvertebrates and fish were
signifi-cantly different over time, largely as a function of
season(lowest densities in early spring, greatest in
summer).Macroinvertebrate, but not fish, assemblage composi-tion
was different by time but not riffle. Although hy-drogeomorphic
characteristics (e.g., streamflow veloci-ty, substrate size) were
linked to shifts in macroinverte-brates and fish, chemical
water-quality parameters (e.g.,dissolved oxygen, nutrient
concentrations) were alsoimplicated as potential biotic drivers.
Our results indi-cate that riffle habitat development can be an
importantmechanism related to restoring sensitive species
andbiological diversity following dam removal.
Keywords Aquaticbiodiversity .Damremoval .Darter .
Hydrogeomorphology. River restoration
Introduction
Riffles—shallow sections of streams or rivers with
rapidcurrent—are important habitat units for benthic
macro-invertebrate and fish assemblages (Kessler and Thorp1993;
Kessler et al. 1995; Heino et al. 2004). Rifflesincrease water
turbulence and oxygen concentration andprovide important
microhabitat variability in depth, ve-locity, and substrate for
aquatic macroinvertebrates(Statzner et al. 1988; Merritt et al.
2008). The highsubstrate heterogeneity typical of riffles (Gordon
et al.2004) also allows multiple fish species to coexistthrough
spatial partitioning of habitat (Kessler andThorp 1993). Likewise,
riffles can support higher den-sities of benthic macroinvertebrates
than pool habitatsand are important areas of food production for
insectiv-orous fishes (Scullion et al. 1982; Gordon et al.
2004).
Even relatively small river infrastructure, such asweirs and
run-of-river, lowhead dams (< 7.6 m inheight) impound water and
can cause a general flat-tening of the channel and fining of
streambed sub-strate, leading to a loss of riffles and gravel
substratesupstream of the structure (Doyle et al. 2005; Salantet
al. 2012). However, in recent years, increasingnumbers of dams have
been removed owing to failinginfrastructure, impounded sediment,
danger posed tohumans, or general lack of utility (Bednarek
2001;American Rivers 2015).
Environ Monit Assess (2018) 190:
339https://doi.org/10.1007/s10661-018-6716-1
Electronic supplementary material The online version of
thisarticle (https://doi.org/10.1007/s10661-018-6716-1)
containssupplementary material, which is available to authorized
users.
D. R. Cook : S. M. P. Sullivan (*)Schiermeier Olentangy River
Wetland Research Park, School ofEnvironment and Natural Resources,
The Ohio State University,Columbus, OH 43202, USAe-mail:
[email protected]
http://orcid.org/0000-0003-2341-5316http://crossmark.crossref.org/dialog/?doi=10.1007/s10661-018-6716-1&domain=pdfhttps://doi.org/10.1007/s10661-018-6716-1
-
Channel responses to dam removal vary considerablydepending on
channel characteristics and sediment re-gimes (Doyle et al. 2003;
Wildman and MacBroom2005), and stem from complex adjustment
processesbetween channel aggradation and degradation(Bushaw-Newton
et al. 2002; Cooper 2013). Responsesto dam removal can be immediate
to longer term (Hartet al. 2002; Doyle et al. 2003), although
increasingevidence suggests that many river responses can
occurwithin months, not years (Grant and Lewis 2015).Maloney et al.
(2008) found that bed particle size bothabove and below the dam
site increased within 1 yearfollowing the removal of a lowhead dam
(105 m wideand 1.7 m in height) on the Fox River, Illinois,
USA.This finding runs counter to other studies that havedocumented
a short-term decrease in particle size atdownstream reaches
following dam removal (e.g.,Thomson et al. 2005). In some
gravel-bed rivers, thetransport rate of sediment exceeds the
sediment- supplyrate; thus, all but the coarsest material was
rapidlyremoved in the former impounded areas 5 years follow-ing
four lowhead dam removals in Connecticut, USA(Wildman and MacBroom
2005). Finer-grained sedi-ments flushed downstream can expose
riffles in theformer impoundment (Egan 2001).
Conversely,courser-grained riffles can be buried by
finer-grainedsediment being transported to downstream reaches
fol-lowing dam removal (Pizzuto 2002). Marked changes inchannel
gradient can also occur via the development ofknick points (Schumm
et al. 1984; Doyle et al. 2003).
The redevelopment of riffles following dam removalmay be an
important factor related to the effects of damremoval on riverine
biotic communities (e.g., Sullivanand Manning 2017). For instance,
an initial increase inmacroinvertebrate abundance has been observed
up-stream of previous dam sites (Bushaw-Newton et al.2002; Maloney
et al. 2008) in contrast to declines inabundance downstream
(Thomson et al. 2005).Maloney et al. (2008) found that the relative
abundanceof Ephemeroptera, Plecoptera, and Trichoptera
(EPT)increased—largely due to increased hydropsychidcaddisfly
abundance—in the formerly impounded areafollowing removal of a
lowhead dam. Cooper (2013)observed that while the total number of
macroinverte-brates increased, there was no significant difference
inthe number of EPT families when comparing pre-dam topost-dam
years on the 4th-order Salmon River in Que-bec, Canada.
Macroinvertebrate community responsescan be both rapid (e.g., 2
weeks; Orr et al. 2008), as well
as occur over longer time scales (e.g., 3.5 years;Renöfalt et
al. 2013).
Several studies have shown that fish species richnessand
diversity tend to increase upstream of previous damlocations (e.g.,
Catalano et al. 2007; Ross et al. 2001),returning to lotic-type
communities (Bushaw-Newtonet al. 2002). Conversely, downstream fish
assemblageshave been shown to decline in species richness,
abun-dance, and diversity shortly following dam removal(Catalano et
al. 2007; Gardner et al. 2013). In particular,the potential
redevelopment of riffles following damremoval may be of particular
benefit to aquatic biota.Bushaw-Newton et al. (2002), for instance,
found thatriffle fish species (e.g., Tessellated Darter
[Etheostomaolmstedi], Shield Darter [Percina peltata], and
HogSucker [Hypentelium nigricans]) moved into newlyformed riffles
upstream of a former impoundment 1 yearafter dam removal in a
4th-order stream in southeasternPennsylvania, USA.
As dam removal and subsequent restoration projectsbecome more
common (Pohl 2002; O’Connor et al.2015), understanding how rivers
change followingdam removal is of increasing importance for both
sci-ence and management. In this study, we monitored howpost-dam
removal riffle development influenced aquaticbiota. This was not a
before-after study; rather, weassessed the associations between
riffle structure andbenthic macroinvertebrate and fish
assemblages1.5–3 years following the removal of a lowhead damon the
5th-order Olentangy River of central Ohio, USA.We predicted that
riffles that developed in the previ-ously impounded section of the
river (via both in-channel restoration activities and natural
geomorphicprocesses) would be characterized by increased
meansediment particle size, streambed slope, andstreamflow
variability over time with concomitantincreases in the density and
diversity of both benthicmacroinvertebrate and fish assemblages
(althoughseasonal variation was expected; e.g., Sullivan andManning
2017). We also hypothesized that dissimi-larities in species
composition and mean abundanceof macroinvertebrates and fish above
and below theprevious dam would decrease over time. Althoughour
focus was on hydrogeomorphic-biotic relation-ships of riffles
following dam removal, we also antic-ipated that chemical water
quality would contribute toexplaining patterns in macroinvertebrate
and fish as-semblages, owing to the importance of water chemis-try
to both fishes and macroinvertebrates (Wynes and
339 Page 2 of 14 Environ Monit Assess (2018) 190: 339
-
Wissing 1981; Rosenberg and Resh 1993; Heringet al. 2006).
Materials and methods
Study system and experimental design
The study area was a 3-river kilometer (rkm) section ofthe lower
Olentangy River, a tributary of the SciotoRiver in central Ohio
(Fig. 1). The Olentangy River isa mixed-bed river, comprised mostly
of gravel andcobble. The 5th Avenue dam (143.3 m wide and 2.5
mhigh) was removed in late summer 2012 in order toimprove water
quality and aquatic habitat. Restorationefforts associated with the
dam removal included chan-nel restoration at sections of a 2.6-km
segment upstreamof the previous dam. Two of our study riffles
(riffles 2
and 3, Fig. 1) were included in an actively restoredsection,
where natural channel design was used to nar-row channel width,
plant riparian vegetation, and rede-velop and reconnect floodplain
wetlands allowing theriver to regain the more natural form and
functions thatexisted pre-dam construction (see Ohio EPA 2011
foradditional details). Among other objectives, goals ofchannel
restoration activities included achievingWarmwater Habitat
designation for fish (as defined bythe Ohio EPA), increasing both
fish and macroinverte-brate community diversity, and meeting the
classifica-tion of a C4 channel as described in Rosgen
(1994:riffle-pool sequence with predominantly gravel sub-strate,
gentle slope, and point bars with well-definedfloodplain). Before
dam removal, no riffles were presentin the impounded sections of
the river (i.e., the areaaligning to location of riffles 2–6
post-dam removal;Fig. 1). Sediments in the impounded area upstream
ofprevious dam were poorly sorted and consisted of sand(58%), ≥
gravel (35%), and silt/clay (7%) (Stantec2010). Before dam removal,
riffle 7 (downstream ofthe previous dam; Fig. 1) was well developed
withpredominantly gravel and cobble substrate.
In total, we surveyed five riffles upstream of theprevious dam
location, one riffle below the previousdam location, and one riffle
downstream of an existinglowhead dam in the same river, which
served as areference site (Fig. 1). Within each riffle, three
quadratswere established at the upstream, middle, and down-stream
portions of the riffle to characterize representa-tive
microhabitats based on flow and substrate charac-teristics. Fish,
benthic macroinvertebrate, chemical wa-ter-quality, substrate, and
streamflow surveys were con-ducted within each quadrat at six time
intervals: latespring (June), summer (August), and late fall of
2014;early spring (March), late spring (June), and summer(August)
of 2015.
This is not a before-after study; in fact, riffles
werenon-existent (or obscured) in the impounded area beforedam
removal, making comparable benthic macroinverte-brate collections
not feasible at riffles 2–6. However, wecollected
macroinvertebrates downstream of the dam be-fore removal (riffle
7). We also include benthic macroin-vertebrate data from the Ohio
Environmental ProtectionAgency from both before and after dam
removal for bothour reference site (Ohio EPA 1999, 2005; Mike
Bolton,Ohio EPA personal communication) (riffle 1; Fig. 1).These
Bbefore^ data were used as points of referenceonly and were not
included in subsequent analyses.
Fig. 1 Sample riffles (filled circles) upstream and downstream
ofthe previous 5th Avenue dam, Columbus, Ohio. Five riffles
devel-oped since dam removal (i.e., riffles 2 through 6). Riffle 1
wasimmediately downstream of an existing lowhead dam of similarsize
and age to the removed 5th Avenue dam and was sufficientlyfar
upstream of the previous 5th Avenue dam to be
considerednon-impounded (free-flowing) (Stantec 2010). Thus, this
riffleserved as our reference reach. Riffles 2 and 3 were located
in anactively restored section of the river following dam removal
(i.e.,natural channel design). Riffle 7 also existed before dam
removal
Environ Monit Assess (2018) 190: 339 Page 3 of 14 339
-
Physical habitat
For each quadrat at each sampling period, Wolman’s(1954)
pebble-count method was used to estimate bedgrain-size distribution
with 200 haphazardly selectedclasts measured using a gravelometer
(e.g., D50 = particlesize for which 50% of the particles are
finer). Size cate-gories (diameter) were as follows: fines (<
0.25 mm),sand (0.25–2.0 mm), gravel (2–16 mm), pebble (16–64 mm),
cobble (64–256 mm), boulder (> 256 mm),and bedrock (solid
surface). Relative grain roughnessfor each quadrat was calculated
using the ratio ofstreamflow depth divided by D84 (i.e., 84th
percentile ofsediment distribution). Streamflow velocity (m s−1)
wasmeasured at two-thirds water depth using a FlowMateModel 2000
(Marsh-McBirney, Loveland, Colorado).Depth (m) was measured with a
stadia rod as distancefrom the water surface to the substrate.
Additionally, atthe beginning and near the end of the study (late
spring of2014 and 2015), two cross-sectional profiles and
onelongitudinal survey of each riffle were conducted todetermine
mean channel slope (m m−1) using a totalstation (Gowin TKS-202,
Beijing, China).
Chemical water quality and nutrients
Temperature, conductivity, dissolved oxygen (DO), andpH were
measured using a YSI 650 MDS® (YSI Inc.,Yellow Springs, Ohio) with
attached 600R® sonde ateach quadrat during each sampling period. In
addition,one 500-ml water sample was collected from each riffle(at
middle of the thalweg) during each sampling periodfor total mercury
(Hg), total nitrogen (N), total phospho-rus (P), nitrate (NO3),
phosphate (PO4), and ammonianitrogen plus phosphate (NH4 + PO4).
The sampleswere stored at 4 °C and sent for analysis at The
OhioState University Service, Testing, and Research
(STAR)Laboratory (Wooster, Ohio), which follows standardmethods and
QA/QC protocols.
Benthic macroinvertebrates and fish
At each quadrat and time interval, a 0.093 m2 (hereafterreported
as 0.10 m2) Surber sampler (500-μm meshsize) was used to collect
benthic macroinvertebrates(90 s per collection) from the stream
bottom followingSullivan and Watzin (2008). Macroinvertebrates
werestored in 70% ethanol and subsequently sorted from
substrate material, identified to family using Merrittet al.
(2008) as a guide, and enumerated.
Fish assemblages were surveyed within each quadratat each time
interval using a Smith-Root® LR-24 (Van-couver, Washington)
backpack electrofisher under nor-mal flow conditions. To prevent
fish from leaving thequadrat, a frame with a weighted net (4.76-mm
mesh)was deployed around the edge of the quadrat (modifiedfrom Bain
et al. 1985). Pulling on upstream and down-stream release cords
enabled us to remotely set the framenet into final position. A time
delay of 15 min betweensetting the frame and sampling the quadrat
permitted aperiod without disturbance prior to sampling (Bain et
al.1985). One electrofishing pass of 100 s for each quadratwas
conducted (total of 300 s per riffle). After collectionand
following enumeration and identification to species,fish were
released.
Numerical and statistical analysis
Family richness, evenness (J’), Simpson’s Index (1-D),and
density were calculated for benthic macroinverte-brates. Due to low
fish numbers, only species richness,density, and number of darter
species were calculatedfor fish assemblages. These metrics were
calculated foreach quadrat as well as for each riffle for each
timeperiod. Species (or family) evenness (J’) quantifies
therelative abundances of species/families within the as-semblage
and ranges from 0 to 1 where communitieswith an equitability number
closer to 1 represent greaterevenness (Pielou 1975). Simpson’s
Index of Diversity(1-D) also ranges between 0 and 1; values closer
to 1indicate greater sample diversity (Simpson 1949; Pielou1969).
The index represents the probability that twoindividuals randomly
selected from a sample will be-long to different species, or in
this case, family.
Given the spatial distribution of our study sites in thesame
river, we tested for potential spatial autocorrelation(Moran’s I)
among response variables (macroinverte-brate and fish density,
richness, and evenness) and foundno evidence for non-random spatial
patterns (p > 0.05 inall cases) (Moran 1950). All data were
transformed (log[x + 1]) where necessary to meet assumptions of
nor-mality and homogeneity of variance. Linear mixed-effects models
were used to test for differences in fishand benthic
macroinvertebrate metrics among rifflesover time, as well as for
differences in streamflowvelocity, mean water depth, and
sediment-size distribu-tion (D16 and D50) and roughness. Following
Davis
339 Page 4 of 14 Environ Monit Assess (2018) 190: 339
-
et al. (2017), time and riffles (i.e., Bsites^) were includedas
fixed effects; quadrats were nested within study rifflesand
considered a random effect. For water-chemistryparameters, for
which we only had one sample per reachper time period, one-way
analysis of variance (ANOVA)was used to compare all reaches by time
steps. Simpleregression analysis was used to explore potential
rela-tionships between (1) roughness, D50, and changes inchannel
slope and benthic macroinvertebrate and fishmetrics; (2) benthic
macroinvertebrate density and fishdensity. Mixed models, regression
analyses, andANOVA were performed using JMP 11.0 (SAS Insti-tute,
Cary, North Carolina).
Non-metric multidimensional scaling (NMS) ordina-tions and
analysis of similarity (ANOSIM; α = 0.05;999 permutations) similar
to Poulos et al. (2014) wereused to examine differences in
macroinvertebrate andfish community composition among sites and
timesteps. NMS enabled visualization of differences in as-semblage
structure among the riffles at the differenttimes and was conducted
separately for benthic macro-invertebrate and fish assemblages
using 500 randomiza-tions and Jaccardian distance matrices (scaled
by vari-ance to provide more equal weight to less
abundantspecies/families), which are generally preferred
forabundance data so that double absences do not contrib-ute toward
distance determination (Legendre andLegendre 1998). In NMS,
distance matrices are rank-ordered and the solution determined
iteratively by min-imization of the stress criterion (Kruskal
1964). TheANOSIM statistic R denotes the magnitude of the
dif-ference among groups; R equals 1 when groups differcompletely
and equals 0 when there is no differencedetected among groups.
Redundancy analysis (RDA) was used to identifypotential
differences in community composition of fishand benthic
macroinvertebrate assemblages as a func-tion of hydrogeomorphic and
water-chemistry predic-tors among riffles (i.e., sites) and over
time. To avoid thenumber of metrics exceeding the number of sites,
theanalysis was limited a priori to four metrics (dissolvedoxygen
[mg L−1], average streamflow velocity [m s−1],D50 [mm], and PO4 [mg
L
−1]). We used these metricsbecause of their importance to
benthic macroinverte-brates and riffle fishes (Kessler and Thorp
1993; Pauland Meyer 2001) and because they reasonably repre-sented
the variability observed in the broader set ofstreamflow and
water-chemistry parameters surveyedas part of this study. Thus, RDA
is useful for
distinguishing the effects of dam removal on macroin-vertebrate
and fish assemblages in terms of driversrelated to either
hydrogeomorphology (average streamflow velocity, D50) or chemical
water quality (DO, PO4).We used R (R Statistical Computing
Software, Vienna,Austria) for ordinations and ANOSIM (R Core
Team2016).
Results
Sediment-size distribution, water depth, and streamflowvelocity
varied by riffle and over time. The density anddiversity of
macroinvertebrates and fish were also dif-ferent over time, largely
as a function of season. Mac-roinvertebrate assemblage composition
was different bytime but not riffle, whereas fish assemblages were
sim-ilar irrespective of time or riffle. Both
hydrogeomorphiccharacteristics and chemical water-quality
parametersemerged as potential drivers of macroinvertebrate andfish
assemblages.
Physicochemical characteristics
Average values (± 1 SD) of water temperature, pH,
andconductivity were 18.5 (± 6.6 °C), 8.40 (± 0.33), and0.615 (±
0.148 μm cm−1), respectively, across the sevenstudy riffles through
all time steps (Supplementary Ma-terial: Table S1). Riffle 7,
downstream of the previousdam, generally exhibited the highest
water temperaturesas compared to the other riffles over time. Total
N andNO3 were lowest in August 2015 and November 2014and highest in
June 2015 (see Supplementary Material:Fig. S1). Total P and PO4
were more consistent overtime, but still exhibited differences
among time periods(note the high PO4 concentration at riffle 1 in
August2014).
Across the study riffles and time periods, gravelranged from
52.7 to 83.0%, and cobble ranged from17.3 to 47.0%. Riffle 2
coarsened the most during thestudy period. The substrate
composition of our referencesite—riffle 1 (downstream riffle of an
existing dam)—remained fairly consistent across the study period.
D16and D50 varied among study riffles and through time(Table 1,
Fig. 2). D16 increased by 8.7 mm fromJune 2014 to June 2015 and by
12.8 mm from August2014 to August 2015 (Supplementary
Material:Table S1). D50 also increased from June and August2014 to
June and August 2015 (Fig. 2). D50 was
Environ Monit Assess (2018) 190: 339 Page 5 of 14 339
-
negatively associated with densities of both Chirono-midae (Fig.
3a) and Hydropsychidae (Fig. 3b).
Streamflow velocity ranged from 0.04 to 1.50 m s−1
(x = 0.50) across all riffles through time and was
signif-icantly different over time (Table 1), which was expect-ed
because of seasonal differences in hydrology (Sup-plementary
Material: Table S1). The lowest flows oc-curred in March 2015, and
the highest occurred inJune 2015 (although note missing values from
Novem-ber 2014). Water depth ranged from 0.04 to 0.85 m (x =0.19 m)
across all riffles and time periods and alsovaried significantly
across time (Table 1, SupplementaryMaterial: Table S1).
Mean channel slope was 0.004 m m−1 across allriffles in June
2014 and 0.008 m m−1 in June 2015.The change in channel slope from
2014 to 2015 (Δslope) was positively related to benthic
macroinverte-brate family richness (Supplementary Material:
Fig.S2a) but negatively related to fish species richness
(Sup-plementaryMaterial: Fig. S2b) and density (Supplemen-tary
Material: Fig. S2c). Relative bed roughness wassignificantly
different through time (Table 1), with thegreatest decrease in
roughness at riffle 4 and the largestincrease in roughness at
riffle 7. We observed no rela-tionship between the change in
roughness and benthicmacroinvertebrate or fish density or diversity
(p > 0.05,data not shown).
Biotic assemblages
Benthic macroinvertebrate density averaged 332.6 indi-viduals
0.1 m−2 across all riffles through time. Macro-invertebrate density
was significantly different amongriffles (Table 1), with riffle 7
exhibiting the greatestdensity (x = 1754.2 individuals 0.1 m−2;
SupplementaryMaterial: Fig. S3). For comparison, density was
2544individuals 0.1 m−2 just prior to dam removal in August
Table 1 Linear mixed-effects models for fish and benthic
macro-invertebrate response variables. BSite^ = study riffle. Also
includ-ed are D16 and D50, relative roughness, streamflow velocity,
andmean water depth
Source df F p
Benthic macroinvertebrates
Density (no. 0.1 m−2)
Site 6, 14 6.24 0.002
Time 5, 70 12.64 < 0.0001
Site*time 30, 70 1.50 0.084
Simpson’s (1-D)
Site 6, 14 1.50 0.249
Time 5, 70 3.36 0.009
Site*time 30, 70 0.82 0.722
Family richness
Site 6, 14 2.16 0.111
Time 5, 70 19.45 < 0.0001
Site*time 30, 70 3.97 < 0.0001
Evenness (J’)
Site 6, 14 1.93 0.146
Time 5, 70 7.45 < 0.0001
Site*time 30, 70 0.79 0.753
Fish
Density (no. 2.25 m−2)
Site 6, 14 1.58 0.225
Time 5, 70 3.47 0.007
Site*time 30, 70 0.51 0.976
Species richness (S)
Site 6, 14 1.24 0.345
Time 5, 70 3.80 0.004
Site*time 30, 70 0.52 0.974
Darter species richness
Site 6, 14 0.93 0.501
Time 5, 70 3.89 0.004
Site*time 30, 70 0.74 0.820
Hydrogeomorphology
D16 (mm)
Site 6, 14 3.05 0.040
Time 5, 70 31.67 < 0.0001
Site*time 30, 70 2.05 0.007
D50 (mm)
Site 6, 14 7.08 0.001
Time 5, 70 12.44 < 0.0001
Site*time 30, 70 0.76 0.791
Relative roughness
Site 6, 14 2.64 0.063
Time 4, 56 3.51 0.013
Table 1 (continued)
Source df F p
Site*time 24, 56 1.18 0.295
Streamflow velocity (m s−1)
Site 6, 14 1.50 0.219
Site*time 24, 56 1.31 0.203
Average depth (m)
Site 6, 14 1.40 0.251
Time 4, 56 9.08 < 0.0001
Site*time 24, 56 0.76 0.756
339 Page 6 of 14 Environ Monit Assess (2018) 190: 339
-
2012. Macroinvertebrate density was also different overtime
(Table 1; Fig. 4a). As a point of reference, macro-invertebrate
abundance at riffle 1 (as assessed by theOhio EPA) was variable
between 1987 and 2011, andthe abundance in 2015 was within the
previous range
and on par with 1999 and 2004 (Supplementary Mate-rial: Fig.
S4).
Fish density across all riffles at all time steps aver-aged 0.4
individuals 2.25 m−2 (SupplementaryMaterial:Table S1). Fish density
was also significantly differentby time (Fig. 4b). Fish density was
not different amongstudy riffles (Table 1). There was no
relationship be-tween benthic macroinvertebrate density and fish
den-sity at any of the study riffles or time steps (p >
0.05,data not shown).
Benthic macroinvertebrate family richness rangedfrom 3 to 17 (x
= 10.8) across all riffles through time(Supplementary Material:
Table S1). Thirty insect fam-ilies were represented in 12 orders,
as well as classOligochaeta and phylum Platyhelminthes. The
mostabundant families were Hydropsychidae, Chirono-midae, and
Baetidae, comprising 41, 21, and 11% ofthe total number of
individuals collected over the study,respectively. Benthic
macroinvertebrate richness waslowest in March across the year, but
was comparableacross the study riffles in June 2014 and June 2015
(Fig.4c). Across six of the seven sites (except riffle 1), therewas
a decrease in macroinvertebrate density from Au-gust 2014 to August
2015. Macroinvertebrate richnesswas not different among study
riffles, but was signifi-cantly different through time (Table 1).
Linear mixedmodels also indicated a significant interaction
effectbetween time and study riffle (p < 0.0001) for
macroin-vertebrate richness, with the effect of time greater
forriffles 1, 2, and 5 than for the others.
Macroinvertebrateevenness also varied significantly over time
(Supple-mentary Material: Table S1), showing greatest evennessat
riffle 4 and lowest at riffle 6 (data not shown).Simpson’s Index
for macroinvertebrates was not signif-icantly different among study
riffles over time (Table 1).
a
a
a
b bb
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
June 2014 August
2014
November
2014
March
2015
June 2015 August
2015
D50
(mm
)
Riffle 1
Riffle 2
Riffle 3
Riffle 4
Riffle 5
Riffle 6
Riffle 7
Fig. 2 Median sediment size, D50, by study riffle over
time(F5,70 = 12.44, p < 0.0001). Riffle 1 is the reference
riffle, riffle2–6 are upstream (of previous dam), and riffle 7 is
downstream ofthe previous dam. Significant pairwise differences are
indicated by
different letters a and b (Tukey’s HSD: p < 0.05). For visual
clarity,error bars are not included. However, please see
SupplementaryMaterial: Table S1 for details on data variability
ytisne
Dea
dihc
ysp
ord
yH
(no
. 0
.1 m
-2)
(a)
(b)
D50 (mm)
ytisne
Dea
dim
on
orih
C
(no
. 0
.1 m
-2)
Fig. 3 Relationships between D50 and densities (no. 0.1 m−2) of
a
Chironomidae (y = 406.8–4.87x; R2 = 0.14, p = 0.015) and
bHydropsychidae (y = 802.9–9.95x, R2 = 0.16, p = 0.008).
Dashedlines represent confidence curves at α = 0.05
Environ Monit Assess (2018) 190: 339 Page 7 of 14 339
-
Fish species richness and darter species richness bothranged
from 0 to 3 by study riffle. Richness of all fishspecies (Fig. 4d)
and darter species (p = 0.004; data notshown) varied temporally
across the riffles, but notamong study riffles (Table 1). The most
common darterspecies included Banded (Etheostoma zonale) and
Rainbow Darters (E. caeruleum), occurring in five andsix of the
study riffles, respectively. Rainbow and Band-ed Darters were found
at all study sites except for riffle7. Greenside Darters (E.
blennioides) were only found atriffle 1. Bluebreast Darters (E.
camurum) were onlyobserved in August 2015 at riffles 2 and 4.
Neither
(a)
(b)
a
b
ab
c
a
a
0
500
1000
1500
2000
2500
3000
3500
etar
betrev
niorc
aM
ciht
neB
Den
sity
(no
. 0
.1 m
-2)
ab
ab
ab
a
b
ab
0
1
2
3
4
5
6
7
m5
2.2
.o
n(ytis
neD
hsiF
2)
(c)
(d)
ab
bc
c
a
bc ab
0
5
10
15
20
25
etar
betrev
nio
rca
Mci
htne
B
Fa
mil
y
Ric
hn
ess
ab ab ab
a
b ab
0
1
1
2
2
3
3
4
June 2014 August
2014
November
2014
March 2015 June 2015 August
2015
ssen
hciR
seicep
Shsi
F
Time
Riffle 1
Riffle 2
Riffle 3
Riffle 4
Riffle 5
Riffle 6
Riffle 7
Fig. 4 Benthicmacroinvertebrate and fishassemblage responses by
studyriffle over time. a Benthicmacroinvertebrate mean density(no.
0.1 m−2) (F5,70 = 12.64,p < 0.0001). b Fish density (no.2.25
m−2) (F5,70 = 3.47, p =0.007). c Benthicmacroinvertebrate family
richness(F5,70 = 19.45, p < 0.0001). d Fishspecies richness
(F5,70 = 0.52, p =0.975). Riffle 1 is the referenceriffle, riffles
2–6 are upstream (ofprevious dam), and riffle 7 isdownstream of the
previous dam.Significant pairwise differencesbased are indicated by
differentletters a, b, and c (Tukey’s HSD:p < 0.05). For visual
clarity, errorbars are not included. However,please see
SupplementaryMaterial: Table S1 and Fig. S3 fordetails on data
variability
339 Page 8 of 14 Environ Monit Assess (2018) 190: 339
-
Simpson’s Index nor evenness of fish assemblages var-ied among
riffles or over time (Table 1).
Shifts in assemblage structure
NMS ordination separated benthic invertebrate assem-blage
composition by time but not by study riffle alongNMS1 (axis 1),
which was confirmed by analysis ofsimilarity (ANOSIM—time: R =
0.588, p = 0.001;site/riffle: R = 0.012, p = 0.371) (Fig. 5a). For
fish as-semblages, there was no difference in composition bytime
(Fig. 5b), or by site (time: R = 0.026, p = 0.347;site/riffle: R =−
0.028, p = 0.570).
Across all study riffles, there was a large proportionof
unconstrained variance identified by RDA. For spe-cies and site
scores, species were scaled proportionallyto associated
eigenvalues, while sites remainedunscaled. Across all study riffles
over time, RDAshowed that benthic macroinvertebrate abundance hada
weak positive association with D50 (data not shown).PO4 was at
least weakly (and positively) related tobenthic macroinvertebrate
density in all but riffles 1and 5 (mostly in late spring and
summer; see Supple-mentary Material: Fig. S5). DO positively
aligned withbenthic macroinvertebrate family richness (riffles 1,
2,and 6; see Supplementary Material: Fig. S5a, b, f) butnot density
(but note negative relationship at riffle 5;Supplementary Material:
Fig. S5e) for certain time pe-riods. Of the physicochemical
variables, streamflowvelocity emerged as the most influential
physicochemi-cal variable from the RDA, where it aligned with
mac-roinvertebrate family richness or density at multipleriffles,
but was inconsistent relative to the nature of theassociation
(Supplementary Material: Fig. S5).
Although we detected no differences in fish assem-blages by time
or riffle, fish and darter species richnesswere positively
associated with D16, D50, andstreamflow velocity at multiple
riffles and time periods(e.g., riffles 1, 2, 3, 5, 7); fish metrics
were negatively(but weakly) related to DO (Fig. 6).
Discussion
Riffle development following dam removal
Altered streamflow patterns characteristic of impoundedrivers
destroy riffles within the reservoir pool and im-pede the
maintenance of riffles both further upstream as
Fig. 5 Non-metric multidimensional scaling (NMS) ordinationplots
of a benthic macroinvertebrate and b fish assemblage com-positions
grouped by date (scaled by variance). The stress valueswere 0.22
and 0.08, respectively. The different shapes indicate thedifferent
riffles and the different colors indicate the different sam-pling
time periods; only the most significant families are
indicated.Dates: June 2014 = red, August 2014 = yellow, November
2014 =green, March 2015 = cyan, June 2015 = gray, August 2015 =
blue.Note that there are differences in the temporal representation
offish data in b (i.e., no fish were found in March 2015 and
August2015); otherwise, colors of the polygons are as noted in a.
Rifflesare shown by symbol: riffle 1 = circle, riffle 2 = square,
riffle 3 =diamond, riffle 4 = triangle (up), riffle 5 = triangle
(down), riffle6 = asterisk, riffle 7 = plus. Dtr. darter
Environ Monit Assess (2018) 190: 339 Page 9 of 14 339
-
well as downstream. After dam removal, new riffles canform in
previously impounded reaches, although thesebedforms may exhibit
low habitat diversity compared toreference riffles (Burroughs et
al. 2009). We observedthat five riffles had developed upstream of
the previousdam (in the former pool and its tail; riffles 2–6). Two
ofthe upstream reaches were within an area of activechannel
restoration, likely leading to rapid development.Similar to other
studies (e.g., following removal of ~2 m high Manatawny Creek Dam,
Pennsylvania;Bushaw-Newton et al. 2002), we also documented
acoarsening of bed sediment: average gravel compositiondecreased
through time from 70.4 to 60.3% across allriffles, while average
cobble composition increasedfrom 29.4 to 39%. We observed no shift
in grain-sizedistribution at the reference site (riffle 1).
Increases in average substrate size in formerlyimpounded areas
are generally in response to higherslope and greater streamflow
velocities (Burroughset al. 2009). In addition to increases in D50,
slope alsoincreased from June 2014 to June 2015 across the
riffles(although note there was substantial variability
amongsites). Similar to findings by Cooper (2013), variabilityin
streamflow velocity was also likely a key driver ofriffle
development in our system, and the interactionsamong these
variables are potentially critical.
The six upstream riffles (riffles 1–6) exhibited greatermedian
particle size (D50) than the downstream riffle
(riffle 7) (but note that riffle 7 increased through time
aswell). This finding is similar to that reported by Thom-son et
al. (2005), who found that particle size wasreduced in the
downstream riffles after the ManatawnyCreek Dam removal.
Conversely, Kanehl et al. (1997)reported that mean percentage rocky
substrate increasedin the formerly impounded areas through time (~4
years) following removal of a lowhead dam on theMilwaukee River,
yet sediment size (reported as theaverage across the study site)
remained similar over timedownstream of the previous dam site.
Thus, smallerparticle size in our downstream reach (riffle 7)
supportsour predictions and suggests at least some degree
ofdownstream transport of fine sediment from the previ-ous
impoundment, despite evidence that the transport ofsediment over
lowhead dams is not fully restricted ow-ing to their small size
(e.g., Stanley et al. 2002; Tulloset al. 2014).
Benthic macroinvertebrate and fish assemblages
Our results were similar to those of Poulos et al. (2014),who
documented changes in biotic communities in bothriffles upstream
and downstream of the previouslowhead dam. As anticipated, we
observed strong sea-sonal changes in macroinvertebrate density and
richness(Fig. 4a, c). Macroinvertebrate density, but not richnessor
assemblage structure, varied by study riffle with riffle
Legend
A: Riffle 2, 2014-06-01; Riffle 3, 2014-06-01; Riffle 6,
2014-06-01; Riffle 7, 2014-06-01; Riffle 5, 2014-08-01; Riffle 6,
2014-08-01; Riffle 2, 2014-11-01; Riffle 4, 2014-11-01; Riffle 6,
2014-11-01; Riffle 7, 2014-11-01; Riffle 1, 2015-03-01; Riffle 2,
2015-03-01; Riffle 3, 2015-03-01; Riffle 4, 2015-03-01; Riffle 5,
2015-03-01; Riffle 6, 2015-03-01; Riffle 7, 2015-03-01B: Riffle 1,
2014-06-01; Riffle 7, 2014-08-01; Riffle 1, 2014-11-01; Riffle 7,
2015-06-01C: Riffle 4, 2014-08-01; Riffle 2, 2015-06-01D: Riffle 4,
2014-06-01; Riffle 5, 2014-06-01; Riffle 1, 2014-08-01; Riffle 3,
2014-08-01; Riffle 3, 2014-11-01; Riffle 5, 2015-06-01; Riffle 1,
2015-08-01; Riffle 5, 2015-08-01; Riffle 7, 2015-08-01 E: Riffle 5,
2014-11-01; Riffle 1, 2015-06-01; Riffle 3, 2015-08-01; Riffle 4,
2015-08-01
Fig. 6 Redundancy analysis(RDA) of fish density (fish total),and
darter species richness(Darter spp) at each study riffleacross all
the time steps. Bluearrows indicate howenvironmental variables
wereordinated. Overall, there were 42riffle/date combinations but
only9 unique scores. Where possible,riffles and time steps (e.g.,
Riffle3, 2015-06-01) are included in theplot. For locations where
labelswould overlap and be illegible,letters A, B, C, D, and E
representthe respective riffles and timesteps as listed in the
legend
339 Page 10 of 14 Environ Monit Assess (2018) 190: 339
-
7 generally exhibiting the highest densities throughtime.
However, density at this site still was lower thanjust before dam
removal and potentially linked to anoverall fining of the sediment
in this reach followingdam removal. Fine sediments commonly
increaseembeddedness (percent of fine sediment surroundinglarge
gravel and cobbles), which can reduce benthic-insect abundances
(Nerbonne and Vondracek 2001) andis a likely mechanism for the
lower macroinvertebratedensity at riffle 7 following dam removal.
For compar-ison, the dominant taxa in impounded areas before
damremoval were midges (Chironomidae) and aquaticwo rms , whe r e a
s f o l l ow i ng d am r emova lHydropsychidae, Chironomidae, and
Baetidae repre-sented the core macroinvertebrate community
(Stantec2010 and references therein). Maloney et al. (2008)
alsoreport a shift from impounded to free-flowing
macroin-vertebrate assemblages following removal of the
SouthBatavia dam (105mwide, 2.7 m high) on the Fox
River,Illinois.
Overall, we found limited evidence to support ourhypotheses that
hydrogeomorphic characteristics wouldrelate to macroinvertebrate
density and diversity, despitesubstantial evidence for links
between stream hydrogeo-morphic features and macroinvertebrates
(Sullivan et al.2004; Bey and Sullivan 2015; Friberg et al.
2009).Across the study reaches, D50 was negatively relatedto
specific families of insects (e.g., Chironomidae),which is
consistent with known tolerance of Chiron-omidae assemblages to
sediment pollution (Zweigand Rabeni 2001; Carew et al. 2007, but
note thatspecific chironomid taxa may be more sensitive toshifts in
sediment). We also observed associationsbetween macroinvertebrates
and both substrate sizeand streamflow velocity (from the RDA), but
thevariable nature of the relationships (i.e., a mixture ofpositive
and negative associations with macroinver-tebrate density or family
richness) makes interpreta-tion difficult. Associations between
water chemistry/nutrients and macroinvertebrate communities
werestrongest for PO4 (and limited evidence for DO; Sup-plementary
Material: Fig. S5). Within the concentra-tion ranges observed in
this study, positive relation-ships observed between PO4 and
macroinvertebratedensity may be a result of increased grazing
opportu-nities without leading to toxicity or eutrophicationlevels.
Additional associations might have been ex-pected if the range of
water-chemistry values hadapproached toxicity thresholds.
Fish density increased significantly from June 2014to June 2015
in our study, even though overall numberswere relatively low
compared to riffles in other systemsin the area (e.g., 1.5
individuals m−2 in Big Darby Creek,Ohio [Bey and Sullivan 2015] vs.
0.5 individuals m−2 inthis study). Bushaw-Newton et al. (2002)
found anincrease in riffle fish species downstream of a formerdam,
but we observed no difference in species richnessacross study
riffles. Multiple darter species colonizedthe newly developed
riffles (as well as a few additionalbenthic insectivores, e.g.,
Johnny Darter [Etheostomanigrum] at riffle 4), although darter
richness was con-sistently greatest at our reference site. Schwartz
andHerricks (2007) found that riffle fishes (i.e., dartersand some
cyprinids) remained absent after constructionof pool-riffle
structure in an urban Illinois stream, im-plicating lack of
macroinvertebrate food sources (ratherthan habitat) as the limiting
factor for fish colonization.Although macroinvertebrate density was
related to sub-strate size (Fig. 3) and varied seasonally (Fig.
4c), therewas no difference among upstream study riffles and
norelationship between macroinvertebrate and fish densi-ties. Thus,
in our study, food resource limitation was anunlikely driver of
variability in riffle fish abundance.
Fish density and darter species richness were relatedto
streamflow velocity (which was relatively consistentacross time,
although there was variability among studyriffles; Supplementary
Material: Table S1) and D16 andD50 in our RDA for multiple sites,
but did not alignstrongly with water-chemistry parameters (Fig. 6).
Thisis in contrast to Hering et al. (2006), who found that
fishresponded more strongly to nutrient enrichment then toland use,
hydrogeomorphology (reach and microhabitatscales), or a
habitat-degradation gradient in a compari-son of 185 streams across
Europe. In addition to thedirect effects of substrate and
variability in streamflowvelocity on both macroinvertebrates and
fishes(Greenberg 1991; Heino 2004; McQuist and Schultz2014),
hydrogeomorphic processes might be expectedto exert indirect
influences on biota via by controllingthe dynamics of dissolved N
and P (Velinsky et al. 2006)and suspended sediment (Kemp et al.
2011).
Although we recorded snapshots of variability instreamflow
velocity and water depth over time, our studyis constrained by the
lack of continuousmeasures of thesevariables, and this will be an
area of important researchrelative to riffle-biota associations
following dam remov-al. Tracking and assessing responses to
high-flow distur-bances may also further illuminate mechanistic
drivers.
Environ Monit Assess (2018) 190: 339 Page 11 of 14 339
-
For example, both macroinvertebrates and fish appearedto respond
to a scouring event in March 2015 withrecovery by June 2015 (Fig.
4), providing initial evidencethat more focused research on the
impacts of critical flowevents on linked geomorphic-biotic
responses followingdam removal may be warranted.
Conclusions
We observed differences in both benthicmacroinvertebrateand fish
assemblages among riffles and over the 15monthsof this study, with
the strongest gains in diversity anddensity in riffle fish
assemblages. Our findings suggest thatriffle formation and the
associated biotic responses follow-ing dam removal are complex and
require consideration ofboth chemical and physical water-quality
characteristics.They also provide initial evidence of the benefits
of rifflehabitat structures as part of dam removal restoration
effortsin gravel-bed rivers, supporting the importance of
rifflemorphology for aquatic biodiversity (Brooks et al. 2005;Costa
and Melo 2008; Cianfrani et al. 2009).
Few studies have focused on fine-scale effects ofdam removal
(e.g., riffle habitat units) on linkedphysical-biotic responses,
yet this level of resolutionmay be an important step in further
understanding eco-system responses to lowhead dam removal.
Interdisci-plinary and longer term (> 5 years) monitoring of
eco-logical responses to dam removals in varying habitatsand stream
types will provide a more holistic under-standing of post-dam
removal ecosystem changes. Ad-ditionally, further evaluation of
fish and benthic macro-invertebrate habitat responses to dam
removal will benecessary in order to further inform fish
conservationstrategies as they relate to dam removal.
Acknowledgements We extend our thanks to members of theStream
and River Ecology (STRIVE) Laboratory in the School ofEnvironment
and Natural Resources and to Drs. David Manningand Katie Hossler
for helpful critiques of this manuscript.
Funding informationThis study was funded by NSF DEB-1341215
(SMPS), the
Ohio Department of Natural Resources, Division of
Wildlifethrough the State Wildlife Grants Program and the Ohio
Biodiver-sity Conservation Partnership (SMPS), the Ohio Water
Develop-ment Authority (SMPS), and The Ohio State University.
Compliance with ethical standards
Ethical approval All applicable national and
institutionalguidelines for the care and use of animals were
followed. Allprocedures performed in studies involving animals were
in accor-dance with the ethical standards of The Ohio State
University.
Open Access This article is distributed under the terms of
theCreative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/), which permits
unrestrict-ed use, distribution, and reproduction in any medium,
providedyou give appropriate credit to the original author(s) and
the source,provide a link to the Creative Commons license, and
indicate ifchanges were made.
References
American Rivers. (2015). Improving—or removing—outdated,harmful
dams. 2015. Retrieved from
www.americanrivers.org/initiatives/dams.
Bain, M. B., Finn, J. T., & Booke, H. E. (1985). A
quantitativemethod for sampling riverine microhabitats by
electrofishing.North American Journal of Fisheries Management, 5,
489–493.
Bednarek, A. T. (2001). Undamming rivers: a review of
theecological impacts of dam removal. EnvironmentalManagement, 27,
803–814.
Bey, C. R., & Sullivan, S. M. P. (2015). Associations
betweenstream hydrogeomorphology and co-dependent
mussel-fishassemblages: evidence from an Ohio, USA river
system.Aquatic Conservation: Marine and Freshwater Ecosystems,25,
555–568.
Brooks, A. J., Haeusler, T., Reinfelds, I., & Williams, S.
(2005).Hydraulic microhabitats and the distribution of
macroinver-tebrate assemblages in riffles. Freshwater Biology, 50,
331–344.
Burroughs, B. A., Hayes, D. B., Klomp, K. D., Hansen, J. F.,
&Mistak, J. (2009). Effects of Stronach Dam removal onfluvial
geomorphology in the Pine River, Michigan, UnitedStates.
Geomorphology, 110, 96–107.
Bushaw-Newton, K. L., Hart, D. D., Pizzuto, J. E., Thomson, J.
R.,Egan, J., Ashley, J. T., Johnson, T. E., Horwitz, R. J.,
Keeley,M., Lawrence, J., Charles, D., Gatenby, C., Kreeger, D.
A.,Nightengale, T., Thomas, R. L., & Velinsky, D. J. (2002).
Anintegrative approach towards understanding ecological re-sponses
to dam removal: the Manatawny Creek study.Journal of the American
Water Resources Association, 38,1581–1599.
Carew,M. E., Pettigrove, V., Cox, R. L., &Hoffman, A. A.
(2007).The response of Chironomidae to sediment pollution andother
environmental characteristics in urban wetlands.Freshwater Biology,
52, 2444–2462.
Catalano, M. J., Bozek, M. A., & Pellett, T. D. (2007).
Effects ofdam removal on fish assemblage structure and spatial
distri-butions in the Baraboo River, Wisconsin. North
AmericanJournal of Fisheries Management, 27, 519–530.
Cianfrani, C. M., Sullivan, S. M. P., Hession, W. C.,
&Watzin, M.C. (2009). Mixed stream channel morphologies:
implicationsfor fish community diversity. Aquatic Conservation
Marineand Freshwater Ecosystems, 19, 147–156.
Cooper, J. E. (2013). Effect of dam removal on aquatic
commu-nities in the Salmon, River, New York. Final Report
2013.Cooper Environmental Research, New York, USA.
339 Page 12 of 14 Environ Monit Assess (2018) 190: 339
http://www.americanrivers.org/initiatives/damshttp://www.americanrivers.org/initiatives/dams
-
Costa, S. S., & Melo, A. S. (2008). Beta diversity in
streammacroinvertebrate assemblages: among-site and
among-microhabitat components. Hydrobiologia, 598, 131–138.
Davis, R. P., Sullivan, S. M. P., & Stefanik, K. (2017).
Reductionsin fish-community contamination following lowhead
damremoval linked more to shifts in food-web structure thansediment
pollution. Environmental Pollution, 231, 671–680.
Doyle, M. W., Stanley, E. H., & Harbor, J. M. (2003).
Channeladjustment following two dam removals
inWisconsin.WaterResources Research, 39, 1–15.
Doyle, M. W., Stanley, E. H., Orr, C. H., Selle, A. R., Sethi,
S. A.,& Harbor, J. M. (2005). Stream ecosystem response to
smalldam removal: lessons from the heartland. Geomorphology,71,
227–244.
Egan, J. (2001). Geomorphic effects of dam removal on
theManatawny Creek, Pottstown, PA. Master’s thesis,Department of
Geology, University of Delaware, Newark.
Friberg, N., Sandin, L., & Pedersen, M. L. (2009). Assessing
theeffects of hydromorphological degradation on macroinverte-brate
indicators in rivers: examples, constraints, and outlook.Integrated
Environmental Assessment and Management, 5,86–96.
Gardner, C., Coghlan Jr., S. M., Zydlewski, J., & Saunders,
R.(2013). Distribution and abundance of stream fishes in rela-tion
to barriers: implications for monitoring stream recoveryafter
barrier removal. River Research and Applications, 29,65–78.
Gordon, N. D., McMahon, T. A., Finlayson, B. L., Gippel, C. J.,
&Nathan, R. J. (2004). Stream hydrology: an introduction
forecologists (2nd ed.). West Sussex: John Wiley & Sons
Ltd..
Grant, G. E., & Lewis, S. L. (2015). The remains of the dam:
whathave we learned from 15 years of US dam removals. In G.Lollino,
M. Arattano, M. Rinaldi, O. Giustolisi, J. C.Marechal, & G. E.
Grant (Eds.), River basins, reservoirsedimentation and water
resources (pp. 31–35). New York:Engineering Geology for Society and
Territory Springer.
Greenberg, L. A. (1991). Habitat use and feeding behavior of
13species of benthic stream fishes. Environmental Biology ofFishes,
31, 389–401.
Hart, D. D., Johnson, T. E., Bushaw-Newton, K. L., Howitz, R.
J.,Bednarek, A. T., Charles, D. F., et al. (2002). Dam
removal:challenges and opportunities for ecological research and
riverrestoration. Bioscience, 52, 669–681.
Heino, J., Louhi, P., &Muotka, T. (2004). Identifying the
scales ofvariability in streammacroinvertebrate abundance,
functionalcomposition and assemblage structure. Freshwater
Biology,49, 1230–1239.
Hering, D., Johnson, R. K., Kramm, S., Schmutz, S.,Szoszkiewicz,
K., & Verdonschot, P. F. M. (2006).Assessment of European
streams with diatoms, macrophytes,macroinvertebrates, and fish: a
comparative metric basedanalysis of organism response to stress.
FreshwaterBiology, 51, 1757–1785.
Kanehl, P. D., Lyons, J., & Nelson, J. E. (1997). Changes in
thehabitat and fish community of the Milwaukee River,Wisconsin,
following removal of the Woolen Mills Dam.American Journal of
Fisheries Management, 17, 387–400.
Kemp, P., Sear, D., Collins, A., Naden, P., & Jones, I.
(2011). Theimpacts of fine sediment of riverine fish.
HydrologicalProcesses, 25, 1800–1821.
Kessler, R. K., & Thorp, J. H. (1993). Microhabitat
segregation ofthe threatened Spotted darter (Etheostoma maculatum)
andclosely related Orangefin darter (E. bellum). CanadianJournal of
Fisheries and Aquatic Sciences, 50, 1084–1091.
Kessler, R. K., Casper, A. F., & Weddle, G. K. (1995).
Temporalvariation in microhabitat use and spatial relations in
thebenthic fish community of a stream. American MidlandNaturalist,
134, 361–370.
Kruskal, J. B. (1964). Multidimensional scaling by
optimizinggoodness-of-fit to a nonmetric hypothesis.
Psychometrika,29, 1–27.
Legendre, P., & Legendre, L. (1998). Numerical ecology.
Oxford:Elsevier Science.
Maloney, K. O., Dodd, H. R., Butler, S. E., &Wahl, D. H.
(2008).Changes in macroinvertebrate and fish assemblages in
amedium-sized river following a breach of a low-head dam.Freshwater
Biology, 53, 1055–1068.
McQuist, M. C., & Schultz, R. D. (2014). Effects of
managementlegacies on stream fish and aquatic benthic
macroinvertebrateassemblages. Environmental Management, 54,
449–464.
Merritt, R. W., Cummins, K. W., & Berg, M. B. (2008).
Anintroduction to the aquatic insects of North America.Dubuque:
Kendall Hunt Publishing.
Moran, P. A. P. (1950). Notes on continuous stochastic
phenom-ena. Biometrika, 37, 17–23.
Nerbonne, B. A., &Vondracek, B. (2001). Effects of local
land useon physical habitat, benthic macroinvertebrates, and fish
inthe Whitewater River, Minnesota, USA. EnvironmentalManagement,
28, 87–99.
O'Connor, J. E., Duda, J. J., & Grant, G. E. (2015). 1000
damsdown and counting. Science, 348, 496e497.
Ohio EPA. (1999). Biological and water quality study of
theOlentangy River and selected tributaries 1999—Delawareand
Franklin Counties. OEPATechnical Report MAS/2000-12-6. State of
Ohio Environmental Protection Agency,Division of Surface Water.
Columbus, Ohio.
Ohio EPA. (2005). Biological and water quality study of
theOlentangy River, Whetstone Creek and Select
Tributaries,2003–2004 – Crawford, Delaware, Franklin, Marion,
andMorrow Counties. Ohio EPA Technical Report EAS/2005-12-6. State
of Ohio Environmental Protection Agency,Division of Surface Water.
Columbus, Ohio.
Ohio EPA. (2011). Environmental assessment: 5th Ave dam re-moval
and river restoration. Columbus: Ohio EnvironmentalProtection
Agency.
Orr, C. H., Kroiss, S. J., Rogers, K. L., & Stanley, E. H.
(2008).Downstream benthic responses to small dam removal in
acold-water stream. River Research and Applications,
24,804–822.
Paul, M. J., & Meyer, J. L. (2001). Streams in urban
landscapes.Annual Review of Ecology and Systematics, 32,
333–365.
Pielou, E. C. (1969). An introduction to mathematical
ecology.New York: Wiley.
Pielou, E. C. (1975). Ecological diversity. New York:
Wiley.Pizzuto, J. (2002). Effects of dam removal on river form
and
process. Bioscience, 52, 683–691.Pohl, M. M. (2002). Bringing
down our dams: trends in American
dam removal rationales. Journal of the American WaterResources
Association, 38, 1511–1519.
Poulos, H. M., Miller, K. E., Kraczkowski, M. L., Welchel, A.
W.,Heineman, R., & Chernoff, B. (2014). Fish assemblage
Environ Monit Assess (2018) 190: 339 Page 13 of 14 339
-
response to a small dam removal in the Eightmile Riversystem,
Connecticut, USA. Environmental Management,54, 1090–1101.
R Core Team. (2016). R: a language and environment for
statis-tical computing. Vienna: R Foundation for
StatisticalComputing.
Renöfalt, B. M., Lejon, A. G. C., Jonsson, M., & Nilsson,
C.(2013). Long-term taxon-specific responses of macroinverte-brates
to dam removal in a mid-sized Swedish stream. RiverResearch and
Applications, 29, 1082–1089.
Rosenberg, D. M., & Resh, V. H. (1993). Freshwater
biomonitor-ing and benthic macroinvertebrates. New York: Chapmanand
Hall.
Rosgen, D. L. (1994). A classification of natural rivers.
Catena,22, 169–199.
Ross, S. T., O’Connell, M. T., Patrick, D.M., Latorre, C. A.,
Slack,W. T., Knight, J. G., et al. (2001). Stream erosion and
densi-ties of Etheostoma rubrum (Percina) and associated
riffle-inhabiting fishes: biotic stability in a variable
habitat.Copeia,2001(4), 916–927.
Salant, N. L., Schmidt, J. C., Budy, P., & Wilcock, P. R.
(2012).Unintended consequences of restoration: loss of riffles
andgravel substrates following weir installation. Journal
ofEnvironmental Management, 109, 154–163.
Schumm, S. A., Harvey, M. D., & Watson, C. C. (1984).
Incisedchannels: morphology, dynamics, and control. Littleton:Water
Resources Publication.
Schwartz, J. S., & Herricks, E. E. (2007). Evaluation of
pool-rifflenaturalization structures on habitat complexity and the
fishcommunity in an urban Illinois stream. River Research
andApplications, 23, 451–466.
Scullion, J., Parish, C. A., Morgan, N., & Edwards, R. W.
(1982).Comparison of benthic macroinvertebrate fauna and
substra-tum composition in riffles and pools in the impounded
RiverElan and the unregulated River Wye, mid-Wales.
FreshwaterBiology, 12, 579–595.
Simpson, E. H. (1949). Measurement of diversity. Nature,
163,688.
Stanley, E. H., Luebke, M. A., Doyle, M. W., & Marshall, D.
W.(2002). Short-term changes in channel form and macroinver-tebrate
communities following low-head dam removal.Journal of the North
American Benthological Society, 21,172–187.
Stantec. (2010). 5th Avenue Dam Evaluation–Lower OlentangyRiver
ecosystem restoration project. Columbus: StantecConsulting Services
Inc 212 p.
Statzner, B., Gore, J. A., & Resh, V. A. (1988). Hydraulic
streamecology: observed patterns and potential applications.Journal
of the North American Benthological Society, 7,307–360.
Sullivan, S.M. P., &Manning, D.W. P. (2017). Seasonally
distincttaxonomic and functional shifts in macroinvertebrate
com-munities following dam removal. Peer J, 5, e3189.
Sullivan, S. M. P., & Watzin, M. C. (2008). Relating
streamphysical habitat condition and concordance of biotic
produc-tivity across multiple taxa. Canadian Journal of
Fisheriesand Aquatic Sciences, 65, 2667–2677.
Sullivan, S. M. P., Watzin, M. C., & Hession, W. C.
(2004).Understanding stream geomorphic state in relation to
ecolog-ical integrity: evidence using habitat assessments and
macro-invertebrates. Environmental Management, 34, 669–683.
Thomson, J. R., Hart, D. D., Charles, D. F., Nightengale, T. L.,
&Winter, D. M. (2005). Effect of removal of a small dam
ondownstream macroinvertebrate and algal assemblages in
aPennsylvania stream. Journal of the North AmericanBenthological
Society, 24, 192–207.
Tullos, D. D., Finn, D. S., & Walter, C. (2014). Geomorphic
andecological disturbance and recovery from two small damsand their
removal. PLoS One, 9, e108091.
Velinsky, D. J., Bushaw-Newton, K. L., Johnson, T. E.,
&Kreeger,D. A. (2006). Effects of a dam removal in SE
Pennsylvaniaon stream chemistry. Journal of the North
AmericanBenthological Society, 25, 569–582.
Wildman, L. A. S., & MacBroom, J. G. (2005). The evolution
ofgravel bed channels after dam removal: case study of theAnaconda
and Union City dam removals. Geomorphology,71, 245–262.
Wolman, M. G. (1954). A method of sampling course
riverbedmaterial. Transactions of the American Geophysical
Union,35, 951–956.
Wynes, D. L., &Wissing, T. E. (1981). Effects of water
quality onfish and macroinvertebrate communities of the little
MiamiRiver. The Ohio Journal of Science, 81, 259–267.
Zweig, L. D., & Rabeni, C. F. (2001). Biomonitoring for
depositedsediment using benthic invertebrates: a test on 4
Missouristreams. Journal of the North American
BenthologicalSociety, 20, 643–657.
339 Page 14 of 14 Environ Monit Assess (2018) 190: 339
Associations between riffle development and aquatic biota
following lowhead dam removalAbstractIntroductionMaterials and
methodsStudy system and experimental designPhysical habitatChemical
water quality and nutrientsBenthic macroinvertebrates and
fishNumerical and statistical analysis
ResultsPhysicochemical characteristicsBiotic assemblagesShifts
in assemblage structure
DiscussionRiffle development following dam removalBenthic
macroinvertebrate and fish assemblagesConclusions
References